An Analysis of Load Factors in Generation Power Plants
Total Page：16
File Type：pdf, Size：1020Kb
Does vertical integration have an effect on load factor? – A test on coal-fired plants in England and Wales N° 2006-03 February 2006 José A. LÓPEZ Électricité de France Evens SALIES OFCE Does vertical integration have an effect on load factor? – A test on coal-fired plants in England and Wales * José A. LÓPEZ** Électricité de France Evens SALIES*** OFCE Abstract Today in the British electricity industry, most electricity suppliers hedge a large proportion of their residential customer base requirements by owning their own plant. The non-storability of electricity and the corresponding need for an instantaneous matching of generation and consumption creates a business need for integration. From a sample of half-hour data on load factor for coal-fired power plants in England and Wales, this paper tests the hypothesis that vertical integration with retail businesses affects the extent to which producers utilize their capacity. We also pay attention to this potential effect during periods of peak demand. Keywords: panel data, vertical integration, electricity supply. JEL Classification:C51, L22, Q41. * Acknowledgements: we thank Guillaume Chevillon, Lionel Nesta and Vincent Touzé for their comments and suggestions. The authors are solely responsible for the opinion defended in this paper and errors. ** [email protected]; *** [email protected], corresponding author. Table of contents 1 Introduction ....................................................................................................... p. 1 2 Vertical integration as a natural structure in an industry subject to particularities..................................................................................................... p. 4 2.1 Fragmentation of UK coal-fired electricity generation capacity....................p. 4 2.2 Drivers behind vertical integration.................................................................p. 4 2.2.1 Hedging customers .............................................................................p. 5 2.2.2 Ensuring plant dispatch during peak demand periods........................p. 6 3 Does vertical integration have an effect on load factor?................................ p. 6 3.1 Technical definition of load factor .................................................................p. 6 3.2 Distribution of load factor ..............................................................................p. 7 3.3 Conditional model for load factor ................................................................. p. 8 4 Results................................................................................................................. p. 9 5 Conclusion........................................................................................................ p. 11 Bibliography..................................................................................................... p. 13 Appendices ....................................................................................................... p. 14 Does vertical integration have an effect on load factor? – A test on coal-fired plants in England and Wales 1 Introduction Today, most electricity retailers hedge their supply by owning their own plant. Thus most electricity retailers face a trade-off between the cost of dealing with market uncertainty and the cost of managing production units. The non-storability of electricity and the corresponding need for an instantaneous matching of generation and consumption creates a need for integration (Kahn, 2002: 46). A supplier that owns physical production capacity can bypass the volatile and often illiquid electricity exchanges in order to hedge its customer base. Unlike Industrial & Commercial customers, who often are offered contracts indexed to electricity exchange related prices, residential customers are charged (for the time being) with prices that do not reflect real-time conditions. 1 Electricity is a basic good thus nearly all electricity customers must be given some assurance against energy shortages and rationing. As mentioned in the previous paragraph electricity retailers supply electricity to domestic customers at prices that are not forcibly linked to real-time wholesale prices. Like for any non-storable good, retailers must respond to uncertain demand in the short term by sourcing from capacity that is inflexible in the short-term. 2 The fact that in the short-term to mid-term the electricity supply curve is inelastic can thus lead to significant wholesale price spikes. 3 Note that in general off-peak, prices yield insufficient income to pure generators, which implies that peak profit maximisation strategies may seek to balance off-peak losses (Fitoussi, 2003: 43) The California experience has made it clear that market system design can have negative consequences for consumers and other stakeholders in the well functioning of 1 Furthermore there are logistical constraints on the number of times per year that a large supplier can adjust its prices, on top of regulated notice periods of notice (currently 28 days in the British electricity market). 2 Maintenance, repairs, equipment failures of power plants and primary input price fluctuation (coal in coal-fired plants, etc.) add to this uncertainty. 3 In fact some economists have asked themselves the question, why spikes are not observed more frequently and with more intensity, probably the result of some sort of self-restraint. 1 José A. LÓPEZ, Evens SALIES the physical electricity system. Agents in the electricity sector are dependent on each other since failures in any part of the system can affect other firms’ costs in the short term 4 (Delmas and Tokat, 2003: 8). These constraints are compounded by the fact that in the short term residential demand is not contingent on prices, particularly because of the political and logistic difficulty of confronting them with the wide fluctuation in marginal supply costs on a real-time basis (see Kahn, 2000: 46). The industry structure of the mid-merit coal-fired generation sector in the UK is the result of several mandatory and voluntary plant divestments and subsequent re- concentration. Before 1990 each regional supplier (ex-Regional Electricity Company) was vertically integrated with a distribution network, whilst National Power and Powergen owned most of the thermal generation assets (78%). 5 Plant divestments along with new entry took place, thus reducing both National Power and Powergen’s share of generation capacity and fragmenting the electricity generation industry. Though some market power remains under the New Electricity Trading Arrangement (NETA) (Bower, 2002: 11-12), estimates suggest the “pivotal player strategy” is less often used in the wholesale energy market in England and Wales as a result of the increasing fragmentation of capacity ownership (Frontier Economics, 2003). 6 Plant divestment was accompanied with a rapid move to vertical integration (VI) of generation businesses with supply businesses (once initial regulatory barriers had been lifted). Today’s present market structure is marked by the dominance of the residential supply market by six large vertically integrated firms (albeit the degree of vertical integration varies amongst them, there is a tendency to converge upon a similar business model). Vertically integrated retailers are less exposed to price volatility as they can internally adjust production (supply) to their load (demand) requirements (Delmas and Tokat, 2003: 9) at a lower cost than the one they might face in the wholesale markets. This 4 Notably the failure of ENRON and other energy traders subsequently increased the amounts of collateral requested by most electricity traders from remaining counter parties to cover credit-risk. 5 Whilst nuclear remained in the hands of British Energy and BNFL. 6 The exercise of market power as a result of capacity withholding by multi-plant producers seems to have been more extreme in California than elsewhere (Kahn, 2002: 47). 2 Does vertical integration have an effect on load factor? – A test on coal-fired plants in England and Wales efficiency may be passed on to customers. Among other authors Kahn (2002: 49) believes that continued ownership of generation by suppliers in California would have protected them but they accepted the obligation to freeze their retail rates and subsequently found themselves in great financial penuries as wholesale prices spiked persistently. By monitoring the performance of their generators, integrated retailers can improve supply reliability and accordingly earn returns from alleviating moral hazard problems. It is thus expected that VI is related to the extent to which producers utilize their available capacity, especially in periods of peak demand. Whilst, at the same time VI may raise policy concerns as it increases market concentration and may reinforce the dominant position of an incumbent. The main question raised in this project is whether VI plays a significant role in explaining the distribution of load factor (LF). LF can be defined as the ratio of the output produced by a plant in a certain period and the theoretical maximum that it could have produced. If LFs were higher for vertically integrated firms, more particularly, during peak times then we could conclude that VI has a role to play in limiting gaming in the wholesale electricity production market. Obviously LFs may be influenced by several other determinants such as seasonality, plant vintage, and plant scale. We essentially address these issues